Why having an AI hub won’t make New Zealand rich

20 August 2026

Faculty of Science, Science and technology, AI, Doctoral, Business and economy

The growth of generative artificial intelligence models is fuelling gargantuan investments by Big Tech and Big AI to secure the infrastructure that supports it.

New Zealand has joined a parallel race to attract that capital, presenting its renewable electricity, cool climate and political stability as the foundations of a potential $25 billion-$35b "AI hub".

At the centre of this approach is Prime Minister Christopher Luxon's Invest New Zealand agency. "Modelled on Ireland and Singapore", it is intended to "roll out the welcome mat" for foreign investors and convert offshore capital into "whole ecosystems of growth and innovation".

This strategy appears to rest largely on clearing a path down which it is imagined AI will largely drive itself - that foreign-owned data centre investments will translate into local jobs, innovation and economic development.

While academics and technology leaders have raised serious environmental concerns, the economic case for AI infrastructure - and its contribution to local development - deserves equal scrutiny.

Ireland and Singapore are questionable models to follow, the global AI boom rests on uncertain future demand, and New Zealand's recent experience with Big Tech data centres offers little assurance that vast investment will produce lasting local value.

Misuse of examples

The first is the misuse of Singapore and Ireland as examples of foreign direct investment (FDI) lead development.

Neither became a hub simply by opening the door to foreign capital and loosely facilitating connections.

Singapore's success in capturing value from foreign investment has always rested on extensive state-led investment and equity in key industries and infrastructure.

Through sovereign wealth funds such as Temasek and GIC, the state holds major stakes across banking, aviation, ports and telecommunications, while projects such as Changi Airport and the Jurong Industrial Estate were deliberately built as foundations for its global hub status.

Ireland became a European gateway for Big Tech, enticing Google, Microsoft and Amazon through low taxes, EU membership and access to transatlantic data flows. But more than simply booking profits there, Big Tech built major European operations anchored in Dublin's "Silicon Docks" - areas of cheap land with favourable tax status.

The current New Zealand government's approach appears to borrow its emphasis on geography and investor facilitation without recognition of the scale of public investment and economic coordination that made these places legitimate hubs.

While academics and technology leaders have raised serious
environmental concerns, the economic case for AI infrastructure - and its
contribution to local development - deserves equal scrutiny.

New Zealand's recent experience

To this end, New Zealand's recent experience demonstrates what a foreign-owned cloud infrastructure investment without public control can deliver.
The recent arrival of Big Tech hyperscale data centre infrastructure by the likes of AWS (Amazon Web Services) and Microsoft came with similar promises in the form of multi-billion-dollar investments and claims to supercharge local tech ecosystems.What we now know is that Big Tech significantly reduced its tax burden by routing revenue through its Irish parent company, using inter-company "service fees" to shift hundreds of millions in revenues offshore.

Meanwhile, the promised economic multipliers - jobs, exportable digital industries, global competitiveness - remain largely hypothetical.

Large data centre projects can also reshape the domestic sector in other ways. Smaller local cloud and data centre providers - those who provide the same kinds of storage, software and platform services to government and business - struggle to compete directly with Big Tech services on price and influence.
Instead, they are increasingly pushed toward the physical side of the industry, securing land and power and constructing facilities to support the digital services of multinationals.

Indeed, European investigations into cloud-service ecosystems - from infrastructure and platforms to software and AI - have shown how big-tech dominance constrains local competition and, in turn, local innovation.
The obvious counterpoint is that access to Big Tech services can itself be enabling. Domestic firms gain access to sophisticated computing, software and AI capabilities without having to build them from scratch.

But access is not the same as developing domestic capability. When a handful of foreign firms control the platforms, standards and services on which local companies increasingly depend, they also shape the terms on which those companies compete.

Recent research in the UK and EU has found local firms often find themselves building at the periphery of big-tech proprietary ecosystems, paying rents to access them, and competing with the same providers as they move into adjacent markets.

For nations pushed into predominantly hosting the infrastructure, the resulting imbalance can be sharper still. Local firms provide physical capacity, while the higher-value digital services and strategic decisions remain offshore.
This raises a broader question how the Government's "concierge" strategy would position domestic firms in the emerging AI economy, and whether infrastructure and support roles offer a pathway for local players into the higher-value activities over time.

The current "concierge" strategy risks turning the country into little more than a well-powered refrigerator for the global AI economy.

Datagrid - a warning
These concerns converge in Datagrid's proposed AI data centre near Invercargill.

Datagrid is Singaporean-owned; we know little about its tax practices, and for a multi-billion-dollar investment, it has projected a relatively low number of long-term employees.

Datagrid has also indicated that its customers are likely to be offshore AI and cloud providers, positioning New Zealand less at the centre of an emerging AI economy than at its infrastructrual edge.

Without stronger public stakes, local capability requirements, or other mechanisms to retain value, projects like this are unlikely to make New Zealand an AI hub.

Instead, the current "concierge" strategy risks turning the country into little more than a well-powered refrigerator for the global AI economy - supplying the land, electricity and cooling but receiving far less of the imagined economic spillover.

Where to next?
New Zealand's AI strategy is pulling in two different directions.

On the one hand, the concierge service for foreign investment risks reproducing mistakes of the big-tech cloud markets. On the other, MBIE's AI strategy explicitly prioritises the adoption of existing AI models over their domestic development.

Taken together, these approaches leave New Zealand positioned at both ends as a customer and a host, but not as owner or producer.

For a smaller economy like ours, some reliance on international capital is unavoidable. But foreign investment need not substitute for an industrial strategy.

The irony is that the opportunity is much larger than what Luxon or Invest New Zealand has set out in their investor prospectus.

Across the Tasman, this realisation has come into view through the recent landmark speech by Australian Prime Minister Anthony Albanese on setting his country's AI and data centre agenda.

He moves, at least rhetorically, beyond narrowly treating AI as a question of innovation or adoption. Instead, he has acknowledged the need for an industrial strategy that links across labour, copyright, national security, energy, water and regional development to secure long-term value over a "short-term boom in capital expenditure and construction".

But most consequentially, the speech brings the politics of ownership, control, and public value to the centre of the AI infrastructure debate.

Rather than "subcontracting our [their] sovereignty and security to the control of foreign monopolies", he presents Australia's national broadband network as a model whose continued public ownership has been a means of retaining national control over foundational digital infrastructure even as communications networks become increasingly entangled with private and overseas interests.

None of this means New Zealand should turn away from AI infrastructure or foreign investment. But if the ambition is genuinely to become an AI hub, it must move beyond treating infrastructure as an end in itself.

The government might start by recognising that successful industrial strategies have long involved a more active role for the state, using investment to build domestic capability to retain a greater share of the value created and, where appropriate, maintain forms of public ownership through periods of major technological change.

Angus Dowell is a doctoral student in the School of Environment, Faculty of Science.

This article first appeared in The Post on 14 August as 'If NZ wants to become an AI hub, its approach is no guarantee of future wealth'